These Techniques for Reducing the Size of Routing Tables Stored in Ternary Content - Addressable Memory

To determine how to forward a packet, an Internet router must perform routing lookup on the destination IP address. Since the adoption of classless interdomain routing in 1993, routing an incoming packet requires that the router find the longest routing prefix that matches the destination IP address. Researchers have proposed various software-based schemes to accelerate the lookup function. However, all of these approaches require at least four to six memory accesses. With the requirement for higher throughput, the latency and bandwidth of modern memory architecture severely limit the number of memory accesses a system designer can afford. Clearly, the software-based solutions do not easily scale up to 10-Gbps processing and beyond. Of the several hardware-based solutions proposed, some use dedicated special hardware and others use commercially available content-addressable memory (CAM). CAM allows simultaneous comparison between all indexes and the key (the destination IP address), and the entry corresponding to the matched index can be obtained directly. CAM’s main advantage is that search time is bounded by a single memory access; thus, it can guarantee high lookup throughput. There are two types of CAM: binary, where each bit position stores only 0 or 1, and ternary, where each bit position can store 0, 1, or don’t care. Binary CAM allows only fixed-length comparisons, so it isn’t directly suitable for longest-prefix matching. A possible solution is to store prefixes of varying length in a separate binary CAM, then design external logic to pick the longest matched entry from all matched CAM chips. Ternary CAM (TCAM) could solve the longest-prefix-matching problem more directly. In addition to the index, TCAM also stores a separate mask for each entry. The mask specifies which bits in the index are active, thereby specifying the variable-length prefix. Table 1 is an example routing prefix table stored in TCAM. Commercially available CAM costs much more than conventional memory, and TCAM is even more expensive. In addition, they consume more power and dissipate more heat, posing a system design challenge. Therefore, it would be advantageous to compact the routing table so that the system could use fewer CAM chips. Even for a single CAM chip, a routing table compaction scheme can help reduce power consumption and heat dissipation. Because the number of routing prefixes is increasing steadily, a routing table compaction scheme can also help contain the routing table size explosion. I propose two techniques to compact routing tables stored in TCAM. Huan Liu